ok so i'm a 2025 grad and i got a recruiter reach-out from xAI last month. panicked immediately because there's almost nothing online about what the new grad loop looks like. ended up getting pretty far (made it past the phone screen, no offer in the end) and i want to post what i found so the next person has something to work with.
what i could piece together going in xAI is small, moves fast, and doesn't have a structured new grad program the way Meta or Google do. there's no rotational thing, no dedicated new grad onboarding cohort. you're expected to get productive quickly. this matters for prep because you're not just showing you can do CS fundamentals, you're showing you can own work.
phone screen standard coding question. i got a medium-difficulty graph traversal problem. had to find shortest path with weighted edges and some additional constraints i won't give away exactly. they were watching how i talked through my approach before coding, not just whether i got it. i spent the first 5 minutes walking through examples and edge cases before writing a line of code. that felt right in hindsight.
what i heard from the recruiter about the onsite (i didn't get there) she said it's typically 4 rounds: two coding, one system design (simplified for new grad level), one behavioral. the system design prompt for new grads is supposedly scoped to something like 'design a URL shortener' or similar canonical problems. you want to know the basics: storage estimation, read/write paths, caching, basic scalability.
what i'd prep if i could do it over drill leetcode medium graph, DP, and string problems. not hard, but consistently medium. practice talking out loud while coding. this is actually harder than the coding. read up on distributed systems basics even for new grad. you don't need staff-level depth but you should know what a load balancer does and why you'd cache something. have one or two real stories about projects. not classroom projects if possible. internship work, personal projects with real users, open source contributions.
things specific to xAI as a company they're building Grok and associated infra. if you have any background in LLM inference, training pipelines, or even just know the terminology around transformer architecture, mention it. they're not going to quiz a new grad on it but it signals you're paying attention to what they're actually building.
anybody who's been further in the new grad loop, please drop what you saw. i'm trying to help build out this data.